Agent Runtime Governance is a public seed repository for runtime governance for agents.
Patterns for governing agents during execution: authority, review, receipts, interrupts, and escalation.
- Kill switch: An explicit stop mechanism that halts an agent, workflow, or integration when continued execution is unsafe or unauthorized.
- Circuit breaker: A temporary control that pauses or degrades execution after repeated failures, policy violations, or risk signals until review or recovery completes.
- Guardrail: A runtime constraint that guides acceptable behavior, such as allowed tools, data boundaries, approval requirements, or escalation rules.
- Admission control: The decision point that determines whether a request, tool call, policy packet, or execution step may enter a governed workflow.
- Fail-closed vs fail-open: A fail-closed path denies or stops execution when required governance checks cannot complete; a fail-open path allows execution to continue under those conditions.
- Policy enforcement point: The component that applies a governance decision at runtime, such as blocking a call, requiring review, logging a receipt, or routing to escalation.
Use CONTRIBUTING.md for contribution process and docs/v0.1-boundary.md for the current repository maturity boundary.
Early public seed repository. Candidate namespace only.
- Define the domain clearly.
- Collect prior art and adjacent ecosystem references.
- Provide reusable schemas, examples, and templates.
- Support human review and agent execution.
- Preserve provenance, auditability, and governance boundaries.
- This repo is not a universal standard.
- This repo does not claim ownership of the broader field or terminology.
- This repo does not canonize HUMMBL/BaseN/Ownward concepts unless explicitly marked and audited.
- This repo must not include private, internal, or secret operational content.
All content is exploratory unless later adopted through a reviewed governance path.
This README now defines the core governance terms used by the repository and links first-time contributors to the contribution process and v0.1 boundary.
This repo is part of the HUMMBL AI governance architecture.
| Repo | Purpose |
|---|---|
| hummbl-governance | Governance runtime — kill switch, circuit breaker, cost governor, 34 primitives |
| mcp-server | MCP server for Base120 mental models — 120 reasoning operators for Claude/Cursor |
| base120 | 120 mental models for structured reasoning — stdlib-only Python library |
| crab | CRAB protocol — Check, Reason, Act, Bus for multi-agent turn execution |
| docs | Canonical public documentation — Mintlify-powered |
| hummbl-toolkit | HUMMBL utility toolkit |
Learn more at hummbl.io.
Reuben Bowlby — Founder, HUMMBL
- GitHub: @hummbl-io
- X/Twitter: @ReubenBowlby
- LinkedIn: reuben-bowlby
- Website: hummbl.io